ABSTRACT
The prevalence of carbapenem-resistant Pseudomonas aeruginosa is increasing. Identification of carbapenemase-producing P. aeruginosa will have therapeutic, epidemiological, and infection control implications. This study evaluated the performance of the EDTA-modified carbapenem inactivation method (eCIM) in tandem with the modified carbapenem inactivation method (mCIM) against a large collection of clinical P. aeruginosa isolates (n = 103) to provide clinicians a phenotypic test that not only identifies carbapenemase production but also distinguishes between metallo-β-lactamase and serine-carbapenemase production in P. aeruginosa. The mCIM test was performed according to Clinical and Laboratory Standards Institute guidelines, while the eCIM was conducted as previously described for Enterobacteriaceae. Test performance was compared to the genotypic profile as the reference. mCIM testing successfully categorized 91% (112/123) of P. aeruginosa isolates as carbapenemases or non-carbapenemase producers, with discordant isolates being primarily Guiana extended-spectrum (GES)-type producers. To increase the sensitivity of the mCIM for GES-harboring isolates, a double inoculum, prolonged incubation, or both was evaluated, with each modification improving sensitivity to 100% (12/12). Upon eCIM testing, all Verona integrin-encoded metallo-β-lactamases (VIM; n = 27) and New Delhi metallo-β-lactamases (NDM; n = 13) tested had 100% concordance to their genotypic profiles, whereas all Klebsiella pneumoniae carbapenemase (KPC; n = 8) and GES (n = 12) isolates tested negative, as expected, in the presence of EDTA. The eCIM failed to identify all imipenemase (IMP)-producing (n = 22) and Sao Paulo metallo-β-lactamase (SPM)-producing (n = 14) isolates. KPC-, VIM-, and NDM-producing P. aeruginosa were well defined by the conventional mCIM and eCIM testing methods; additional modifications appear required to differentiate GES-, IMP-, and SPM-producing isolates.
INTRODUCTION
Carbapenem-resistant Pseudomonas aeruginosa (CRPA) is a growing concern globally and accounts for 10 to 50% of P. aeruginosa isolates depending on local epidemiology (1, 2). In the United States, recent data from the U.S. Emerging Infections Program found 10% of P. aeruginosa were carbapenem-resistant (3). Carbapenem resistance in P. aeruginosa is mediated by a variety of mechanisms, including drug efflux, alterations in membrane permeability, and the production of β-lactamases such as serine- and metallo-β-lactamases (MBLs) (4). The prevalence and diversity of MBLs in P. aeruginosa is largely driven by geography, with recent surveillance studies documenting an MBL-producing CRPA prevalence rate of approximately 2% in the United States and as high as 20% in European and Mediterranean countries (5, 6). In addition to Klebsiella pneumoniae carbapenemase (KPC), other serine-β-lactamase-producing subtypes (i.e., Guiana extended spectrum [GES]) are being isolated more frequently in North America, particularly in Canada and Mexico (7–9). Recently, GES-harboring P. aeruginosa isolates were reported in two unrelated patients from Texas (10). With increasing reports of carbapenemase-producing P. aeruginosa, including GES-positive isolates, a simple and reliable method for phenotypic detection is needed. Identification of carbapenemase production can dictate infection control measures in an attempt to curb nosocomial spread (11).
Currently, the Clinical and Laboratory Standards Institute (CLSI) endorses the use of the CarbaNP and modified carbapenem inactivation method (mCIM) for carbapenemase detection in P. aeruginosa (12). However, these methods are limited by need for more expensive materials and in-house preparation of reagents (i.e., CarbaNP) and an inability to distinguish between serine-based enzymes and MBLs (i.e., CarbaNP and mCIM). Other methods for detection, including the MBL-Etest, have been evaluated but found to have low sensitivity for some enzyme subtypes (13). Recently, the EDTA-modified carbapenem inactivation method (eCIM) was validated for the detection of MBL-producing Enterobacteriaceae by using the ability of EDTA to inhibit MBL activity (14). This method is attractive since it utilizes readily available materials and can be run in tandem with the mCIM. In addition, an evaluation of the mCIM and eCIM for the increasingly reported GES-producing isolates is warranted (15). Here, we present data from a study evaluating the utility of the mCIM combined with eCIM testing for the detection of MBL production in a diverse collection of P. aeruginosa isolates.
MATERIALS AND METHODS
Bacterial isolates.Ninety-six carbapenemase-producing P. aeruginosa isolates were evaluated. Of these, 31 isolates were acquired from the Centers for Diseases Control and Food and Drug Administration Resistance Bank (CDC-FDA) Antimicrobial Resistance Isolates Bank, and 65 isolates were from the Center for Anti-Infective Research and Development (CAIRD) isolate repository. All isolates were previously molecularly categorized via PCR or whole-genome sequencing for β-lactamase genes.
Tested isolates harbored a variety of β-lactamase genes. The following MBL enzyme types were represented: Verona integrin-encoded metallo-β-lactamase (VIM; n = 27), imipenemase (IMP; n = 22), New Delhi metallo-β-lactamase (NDM; n = 13), and Sao Paulo metallo-β-lactamase (SPM; n = 14). The following enzyme subtypes were represented in this study: VIM (-2, -5, -4, and -11), IMP (-1, -6, -7, -10, and -48), NDM (-1), and SPM (-1). In addition, 20 class A serine β-lactamase-producing isolates underwent testing: GES (n = 12) and KPC (n = 8). The meropenem MICs were obtained through broth microdilution per CLSI standards (12). Meropenem MICs for all carbapenemase-producing isolates were ≥8 mg/liter. Seven wild-type P. aeruginosa isolates were also tested as negative comparators (meropenem MIC range, 0.125 to 2 mg/liter).
All isolates were stored in skim milk (Becton Dickinson, Sparks, MD) at −80°C and subcultured to Trypticase soy agar with 5% sheep blood plates (Becton Dickinson) prior to incubation. Isolates were incubated at 37°C for 18 to 20 h prior to second subculture (without the use of meropenem disks) before testing.
mCIM/eCIM testing.The mCIM procedure was performed as previously described (12, 13, 16). Briefly, 2-ml aliquots of Trypticase soy broth (TSB) were directly inoculated with a 10-μl loopful of P. aeruginosa colonies. The emulsion was vortexed and a 10-μg meropenem disk (Becton Dickinson, lot 3065664) placed in the inoculated tube. Tubes were incubated for 4 h (±15 min) and placed on a Mueller-Hinton agar plate after lawn inoculation of 0.5 McFarland bacterial suspension of meropenem-susceptible Escherichia coli ATCC 25922. Simultaneous with the mCIM testing, eCIM testing was performed with each isolate (10-μl loopful of P. aeruginosa) as previously described for Enterobacteriaceae (14) and emulsified in a 2-ml aliquot of TSB-EDTA (EDTA concentration, 5 mM). Isolates were incubated and plated as described for mCIM testing. Plates were incubated for 18 to 24 h, and the zone diameter was read by two independent investigators.
The results for both the mCIM and the eCIM were interpreted as previously described (13, 14, 16). The mCIM-positive results were interpreted by measuring the zone of inhibition around the disk inoculated in TSB alone. Results were considered positive if the zone of inhibition was 6 to 15 mm or 16 to 18 mm with pinpoint colonies within the zone. Results were considered indeterminate if the zone was 16 to 18 mm or 19 mm with pinpoint colonies within the zone of inhibition. Negative mCIM results were isolates with a zone of inhibition ≥19 mm in diameter. The eCIM was only interpreted if the mCIM was positive by taking the difference in the zone from the disk in the presence of EDTA minus the zone from the mCIM testing. If the zone diameter from the eCIM minus that from the mCIM was ≥5 mm, the result was considered positive.
The following isolates were used for quality control: Klebsiella pneumoniae ATCC-BAA 1706 (negative for carbapenemase production), K. pneumoniae ATCC-BAA 1705 (KPC positive), K. pneumoniae CDC Bank 505 (NDM positive), and P. aeruginosa ATCC 27853 (negative for carbapenemase production).
Alternative mCIM methods for GES isolates.Previous investigators noted challenges with mCIM identification of GES-producing P. aeruginosa potentially due to the low catalytic activity of this enzyme class (15, 17). To improve mCIM test performance for this class of enzymes, three modifications were evaluated in addition to the standard inoculation (i.e., one 10-μl loopful and 4-h incubation). Thus, a double inoculation (two 10-μl loopfuls), double incubation time (8 h), and double inoculation plus double incubation time were assessed separately with each GES-positive P. aeruginosa. The CLSI interpretive criteria for mCIM were utilized (12). Testing also included a P. aeruginosa isolate negative for carbapenemase production (P. aeruginosa ATCC 27853) and a KPC-producing isolate (CAIRD PSA 1844).
To evaluate any changes in the diagnostic performance using the altered testing method, 10 randomly selected carbapenemase-producing isolates (VIM, n = 4; IMP, n = 3; SPM, n = 2; and KPC, n = 1) from the eCIM analysis and 20 non-carbapenemase-producing isolates harboring various cephalosporinases (Pseudomonas-derived cephalosporinase [PDC]) and/or efflux/porin mutations (meropenem MIC range, 1 to 64 mg/liter) were tested with the traditional mCIM and all three mCIM modifications.
Data analysis.Isolates’ genotypic profiles were used as the reference standard. True-positive results were defined as phenotypic (mCIM or eCIM) concordance with the isolate genotyping. True-negative results were defined as negative phenotypic testing in isolates with a lack of a carbapenemase genotypic profile. False-positive results were defined as a phenotype indicating either carbapenemase production (mCIM) or MBL production (eCIM) in isolates with negative genotypic findings. Finally, false-negative results were defined as negative phenotypic results despite genotypic indication of carbapenemase or MBL enzymes. Discordant results were confirmed by testing up to three times to obtain a mode. Sensitivity and specificity were calculated separately for the mCIM and eCIM testing.
RESULTS
mCIM test.The mCIM successfully categorized the majority (91% [112/123]) of the P. aeruginosa isolates evaluated (Table 1), resulting in a sensitivity of 93% (86 to 97%) and a specificity of 85% (66 to 96%). All VIM-, IMP-, SPM-, and KPC-positive isolates were found to be positive on mCIM testing. Of the 7 false-negative isolates, one harbored an NDM (profile: NDM-1, PDC-35), and the remaining discordant isolates harbored GES-type enzymes. Figure 1 shows a comparison of a positive and negative mCIM results. There was significant result variability within the class of GES-harboring isolates. Initially, GES isolates resulted in four mCIM-positive isolates, four mCIM-negative isolates, and four mCIM-indeterminate isolates. Isolates were retested to obtain a mode ultimately resulting in six mCIM-positive and six mCIM-negative isolates.
mCIM and eCIM results for P. aeruginosa by genotypic profilea
mCIM and eCIM results. (A) Example of an NDM-producing P. aeruginosa with mCIM (left) having growth up to the meropenem disk and eCIM (right) with a clear zone diameter around the disk placement. Lines indicate the zone of inhibition measured for each test. Both result in positive readings for carbapenemase and metallo-β-lactamase production, respectively. (B) KPC-producing P. aeruginosa, with mCIM (left) and eCIM (right) both showing growth up to the meropenem disk. Results interpreted as positive and negative, respectively. The growth around the disk noted in panels B and C represent carryover growth from the P. aeruginosa isolate inoculated with the meropenem disk. (C) Non-carbapenemase-producing P. aeruginosa with negative mCIM (left) and eCIM (right) results that would not be interpreted as the mCIM was negative.
mCIM modification for GES-producing isolates.All three mCIM modifications resulted in positive mCIM tests (100% versus 50% with standard mCIM procedure). Of note, with these mCIM modifications, we observed a halo zone, i.e., a thinning in the lawn of the indicator E. coli around the meropenem disk for all previously mCIM-negative isolates (Fig. 2). This did not impact result interpretation or test specificity. Table 2 shows the GES subtypes for each isolate, the meropenem MIC, and the mCIM testing result for each method.
Halo effect of alternative testing of GES isolates. A GES-positive isolate was tested after doubling inoculum and 8 h of incubation that was previously determined to be negative upon standard mCIM testing. Note the halo around the growth up to the meropenem disk (left). On the right is a positive result from a KPC-producing P. aeruginosa isolate after double the inoculum and a 4-h incubation.
mCIM results for GES enzymes using the standard mCIM and various inocula and incubation periodsa
As expected, all 10 non-GES carbapenemase-producing isolates tested mCIM positive with each modification. Of the 20 non-carbapenemase-producing carbapenem-resistant isolates, four yielded false-positive results with the standard mCIM. When these isolates underwent modified testing, four false positives were noted in the double-inoculum group, while five false positives were observed in both the double-inoculum and double-inoculum-plus-double-incubation groups.
eCIM test.The eCIM diagnostic performance was evaluated in tandem with all mCIM-positive isolates (75 MBL and 14 serine-carbapenemases). All VIM, NDM, KPC, and GES isolates evaluated had 100% concordance to their genotypic profiles (see Table 1). Notably, discordant results were observed among all the IMP- and SPM-producing isolates included in this study, with EDTA unable to inhibit MBL activity. Testing performance results with MBL enzymes are given in Table 3.
eCIM test performance by enzyme type in P. aeruginosaa
DISCUSSION
Routinely utilized for Enterobacteriaceae only, the eCIM testing procedure was able to differentiate VIM and NDM producers from serine-carbapenemase producers in a diverse collection of clinical P. aeruginosa isolates. Due to the ability to test concurrently with the mCIM and use of readily available testing supplies, the eCIM may be a suitable phenotypic method to distinguish between these subtypes of MBL carbapenemase and serine-carbapenemases in P. aeruginosa. Caution should be taken using the current testing conditions in areas with endemic IMP- and SPM-producing P. aeruginosa.
Although MBL production in P. aeruginosa is predominated by VIMs, IMPs are increasingly encountered and, likewise, SPMs are widespread in certain regions, notably Brazil (18, 19). The genotypic-phenotypic discordance observed among IMP-producing P. aeruginosa evaluated in this current eCIM study corroborate previous publications with Enterobacteriaceae. For example, the eCIM with 0.1 mM EDTA resulted in false-negative results for all three IMP-harboring isolates (two Klebsiella pneumoniae isolates and one Klebsiella aerogenes isolate) and two NDM-harboring isolates (one Salmonella isolate and one Proteus mirabilis isolate) evaluated in a recent investigation (14). Upon increasing the EDTA concentration 50-fold (i.e., 5 mM) to improve performance, the same authors observed positive eCIM results, suggesting stronger zinc chelation may be needed to inhibit certain MBL enzymes (i.e., IMP isolates) (14). No SPM-producing organisms were evaluated in the previous eCIM assessment; however, based on our findings, the current eCIM method and EDTA concentration does not appear appropriate to detect SPM-carbapenemase production. Future investigations with eCIM modifications and/or a wide range of EDTA concentrations against IMP- and SPM-producing Enterobacteriaceae and P. aeruginosa to improve the eCIM testing sensitivity are warranted.
Previously characterized as extended-spectrum β-lactamases, some GES variants have displayed carbapenemase activity, including GES-20 (20), which was found in the majority of the GES-positive isolates in our study. GES-type carbapenemases account for the most common carbapenemase detected in P. aeruginosa isolates from Canada, particularly GES-5 (8). GES carbapenemases have also been frequently reported in Mexico, with GES enzymes identified in >90% of imipenem-resistant P. aeruginosa isolates in one multicenter surveillance study (9, 21). This included a high proportion of isolates with tandem GES-19 and GES-20 alleles (9). Previous studies reported challenges in the phenotypic detection of GES-carbapenemase-producing organisms with the CIM and mCIM (15, 22). We observed 6 of 12 GES-positive isolates test negative using the standard mCIM. This finding may be related to decreased expression or variable affinity for meropenem hydrolysis between alleles and/or allele combinations. Lisboa and colleagues described similar variability, as the mCIM failed to identify 4 of 16 GES-5-harboring P. aeruginosa isolates. These investigators hypothesized that due to the lower catalytic activity of GES, increasing the bacterial inoculum may increase the detection yield for the mCIM. Upon retesting with a higher inoculum, the investigators observed an improvement in mCIM test performance (15). Similarly, we sought to evaluate the mCIM test performance on GES-positive isolates by increasing the inoculum, increasing incubation time, or both. We found all three methods produced reproducible, mCIM-positive results. Interestingly, halo zones were found around the meropenem disks for isolates that tested negative with the standard mCIM testing but positive with the modified methods. This may be a result of variable hydrolytic activity as an enzymatic class but did not impact result interpretation or test specificity. Implementation of these modifications may be considered in regions where these GES-positive isolates are common though local validation is needed.
Simner and colleagues evaluated several phenotypic methods for identifying carbapenemase production in 13 carbapenemase-producing and 53 non-carbapenemase-producing P. aeruginosa (13). The authors reported an mCIM test sensitivity and specificity of 100 and 98%, respectively (13). These results were similar to a previous evaluation of 30 P. aeruginosa isolates including 15 carbapenemase-producing strains (23). Our data in a larger collection of clinical P. aeruginosa isolates performed similarly with an mCIM test sensitivity of 93%. Importantly, the strength of this study lies in the diversity of carbapenemase-producing and non-carbapenemase carbapenem-resistant isolates evaluated. Nonetheless, despite retesting in triplicate, the mCIM failed to categorize one NDM-harboring isolate. After reconfirming the presence of the NDM gene in this isolate, the discordant phenotypic test may be attributed to the contribution of other resistance mechanisms resulting in meropenem inactivity (4). This finding suggests leveraging both phenotypic and genotypic detection methods to identify resistance mechanisms to guide therapeutic decisions and infection control efforts (24).
While further data are needed to evaluate methods that improve the test performance for certain enzyme types (i.e., GES, IMP, and SPM), combination mCIM and eCIM testing may be an attractive phenotypic method for identifying carbapenemase-producing P. aeruginosa and may therefore be a valuable tool to the clinical laboratory, especially in low-resource settings where molecular diagnostic testing has limited availability.
ACKNOWLEDGMENTS
We acknowledge the staff at the Center for Anti-Infective Research and Development for assistance with testing.
This project was internally funded by the Center for Anti-Infective Research and Development.
D.P.N. is a consultant for, is on the speakers’ bureau of, or has received research funding from Allergan, Bayer, Cepheid, Merck, Melinta, Pfizer, Wockhardt, Shionogi, and Tetraphase. C.M.G., M.J.L., and T.E.A. have no conflicts of interest to disclose.
FOOTNOTES
- Received 6 December 2019.
- Returned for modification 2 January 2020.
- Accepted 3 February 2020.
- Accepted manuscript posted online 1 April 2020.
- Copyright © 2020 American Society for Microbiology.
